[pulseaudio-discuss] Has anyone ever gotten Orca and Pulse Audio to work together well?
halim.sahin at freenet.de
Sun Dec 13 10:02:58 PST 2009
On Sa, Dez 12, 2009 at 04:32:22 -0500, Daniel Chen wrote:
> On Sat, Dec 12, 2009 at 2:43 PM, Bill Cox <waywardgeek at gmail.com> wrote:
> > I'm trying to narrow down where the problem is that causes the 1/2
> > second delay when Orca tries to read text using pulseaudio. Is it in
> > pulseaudio's main code trunk, or in Ubuntu's version of pulseaudio?
> alsa-plugins, actually.
Not sure, because ubuntu has switched to speech-dispatcher, which has a
native pulse module.
> The issue seems to be that espeak, in Ubuntu, uses an old version of
> PortAudio known not to support PulseAudio adequately. Further, it is
> configured not to use PulseAudio by default but instead routes through
> the pulse pcm & ctl ALSA plugin, which introduces latency. Based on a
> bug report, I'm going to configure it to use the PulseAudio backend
This is true if gnomespeech is in use!
> > Do any of you know of any case where pulseaudio has been configured to
> > deliver speech from Orca without the delay? Is this simply a problem
> > now because Ubuntu was the version most blind users preferred before
> > Ubuntu switched to pulseaudio? Has this problem been there forever?
> This problem has become exacerbated with the advent of PulseAudio
> exposing broken bits in the audio stack: alsa-lib, linux, hardware
> itself. (Note: just because things appeared to work without
> PulseAudio does *not* preclude the brokenness of the other parts in
> the stack!)
I am now using a11y technology since 10 years.
During this time I had never such problems with the soundoutput (with
alsa nor oss).
PA has introduced serious problems for people with disabilities.
1. Per user sessions
This means, a consolescreenreader who needs root access can't use the
This means no feedback of the bootproccess or root account because
the speech server can only be started after pulseaudio.
there is also no feedback of console logins!
2. Configuration highly depends on graphical utilities.
It's not easy to use a console only system.
I have read many mails about not running pulseaudio in the init process.
Unfortunately this would be the ultimate solution for multiuser audio
Back to topic:
It would be great when some pulseaudio developers can look into the
pulseaudio driver of speech-dispatcher.
I think the main problem can be found there.
The mentioned file is located in
Speechd will become the default speech output solution for console and
For gnome it will be the default with 3.0 because of bonobo
More information about the pulseaudio-discuss